It is now possible to design ex novo recognition peptides, called anti-sense peptides based on anti-sense DNA sequence of a gene coding for a target protein. In several cases tested so far, anti- sense peptides have been found to bind to the corresponding sense peptides, with rather profound implications for use in affinity technology to isolate and purify polypeptides of biological relevance. The major limitation for the general application of this new macromolecular recognition system is the lack of DNA sequence information, especially in the case of novel uncharacterized proteins. The main objective of this study is to devise a method to generate recognition peptides using only partial amino acid sequence information. Results obtained so far indicate that it is possible to design recognition peptides starting only from limited sequence information using values of individual amino acid hydropathy. Such ex novo sequence-directed recognition peptides, once immobilized on solid support, maintained their binding properties and proved to be useful in preliminary purification of target proteins.